Обсуждение: Re: pgsql: Improve wording.
Martijn van Oosterhout wrote: > On Mon, Dec 17, 2007 at 01:48:31PM +0000, Alvaro Herrera wrote: > > Log Message: > > ----------- > > Improve wording. > I'd suggest removing everything between the parentheses, or perhaps > something like: By tracking allocated memory rather than used memory > it removes the possibilty of failing in a page fault. This option also > require you allocate enoghh swap to cover all possible eventualities, > though in production it will not be used. I agree that the parenthised phrase should be removed. http://developer.postgresql.org/cvsweb.cgi/pgsql/doc/src/sgml/runtime.sgml?r1=1.387&r2=1.391 I think your proposed wording is removed enough from what the complainant was saying that it is not worth to stick it in. The point here is, to what extent do we want to spoon-feed careless sysadmins? -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
On Tue, Dec 18, 2007 at 09:32:51AM -0300, Alvaro Herrera wrote: > Martijn van Oosterhout wrote: > > On Mon, Dec 17, 2007 at 01:48:31PM +0000, Alvaro Herrera wrote: > > > Log Message: > > > ----------- > > > Improve wording. > > I agree that the parenthised phrase should be removed. > > http://developer.postgresql.org/cvsweb.cgi/pgsql/doc/src/sgml/runtime.sgml?r1=1.387&r2=1.391 That's fine with me. It's just that I figured that if we're going to add something, we should at least add something factually accurate. > I think your proposed wording is removed enough from what the > complainant was saying that it is not worth to stick it in. The point > here is, to what extent do we want to spoon-feed careless sysadmins? Not sure what the complaint was. The only useful advice I can think of is that before turning it on you need to makes sure you have lots of swap. For the rest, I think the wording was fine as it was. Have a nice day, -- Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/ > Those who make peaceful revolution impossible will make violent revolution inevitable. > -- John F Kennedy
Martijn van Oosterhout wrote: -- Start of PGP signed section. > On Tue, Dec 18, 2007 at 09:32:51AM -0300, Alvaro Herrera wrote: > > Martijn van Oosterhout wrote: > > > On Mon, Dec 17, 2007 at 01:48:31PM +0000, Alvaro Herrera wrote: > > > > Log Message: > > > > ----------- > > > > Improve wording. > > > > I agree that the parenthised phrase should be removed. > > > > http://developer.postgresql.org/cvsweb.cgi/pgsql/doc/src/sgml/runtime.sgml?r1=1.387&r2=1.391 > > That's fine with me. It's just that I figured that if we're going to > add something, we should at least add something factually accurate. > > > I think your proposed wording is removed enough from what the > > complainant was saying that it is not worth to stick it in. The point > > here is, to what extent do we want to spoon-feed careless sysadmins? > > Not sure what the complaint was. The only useful advice I can think of > is that before turning it on you need to makes sure you have lots of > swap. For the rest, I think the wording was fine as it was. We are back to the old wording, but a link to an OOM article was added: http://lwn.net/Articles/104179/ -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
Alvaro Herrera wrote: > Martijn van Oosterhout wrote: > > On Mon, Dec 17, 2007 at 01:48:31PM +0000, Alvaro Herrera wrote: > > > Log Message: > > > ----------- > > > Improve wording. > > > I'd suggest removing everything between the parentheses, or perhaps > > something like: By tracking allocated memory rather than used memory > > it removes the possibilty of failing in a page fault. This option also > > require you allocate enoghh swap to cover all possible eventualities, > > though in production it will not be used. > > I agree that the parenthised phrase should be removed. > > http://developer.postgresql.org/cvsweb.cgi/pgsql/doc/src/sgml/runtime.sgml?r1=1.387&r2=1.391 > > I think your proposed wording is removed enough from what the > complainant was saying that it is not worth to stick it in. The point > here is, to what extent do we want to spoon-feed careless sysadmins? OK, I have removed the paratheses paragraph about fork() and added link to an article about the OOM: http://lwn.net/Articles/104179/ On Linux 2.6 and later, an additional measure is to modify the kernel's behavior so that it will not <quote>overcommit</>memory. Although this setting will not prevent the <ulink url="http://lwn.net/Articles/104179/">OOMkiller</> from being invoked altogether, it will lower the chances significantlyand will therefore lead to more robust system behavior. This is done by selecting strict overcommit modevia <command>sysctl</command>: -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +